Quantify the observed rate of failed, missing, stale, or invalid feature reads in an audit sample. Results include errors per 1,000 reads, successful reads, and an approximate confidence interval for the underlying error rate.
The displayed 95% range uses p̂ ± 1.96 × standard error and is clipped to 0%–100%.
Use the main result together with the supporting bounds, counts, or capacity figures. The estimate is only as reliable as the input definitions, sampling process, and operating assumptions.
Define an error consistently before sampling. Combining unrelated failure modes can hide which remediation will have the greatest effect.
24 errors in 1200 audited observations gives an observed error rate of 2.00% and 20.0 errors per 1,000.
Use a written audit rule and count only observations that fail it. Keep the rule consistent across the entire reviewed sample.
A normalized rate makes samples of different sizes easier to compare while preserving the underlying observed percentage.
No. Errors must be between zero and the total number of audited observations.
It is an approximate range for the underlying error proportion if the audited observations form a representative independent sample.
Not necessarily. Confirm that sampling, traffic mix, definitions, and audit difficulty remained comparable.
| Metric | Definition |
|---|---|
| Error rate | Errors divided by audited observations |
| Passing count | Audited observations minus errors |
| 95% interval | Approximate sampling range |